Physiological roles of A₁ and A_2A adenosine receptors in regulating heart rate, body temperature, and locomotion as revealed using knockout mice and caffeine

Abstract

Heart rate (HR), body temperature (Temp), locomotor activity (LA), and oxygen consumption (O₂C) were studied in awake mice lacking one or both of the adenosine A₁ or A_2A receptors (A₁R or A_2AR, respectively) using telemetry and respirometry, before and after caffeine administration. All parameters were lower during day than night and higher in females than males. When compared with wild-type (WT) littermates, HR was higher in male A₁R knockout (A₁RKO) mice but lower in A_2ARKO mice and intermediate in A₁-A_2AR double KO mice. A single dose of an unselective β-blocker (timolol; 1 mg/kg) abolished the HR differences between these genotypes. Deletion of A₁Rs had little effect on Temp, whereas deletion of A_2ARs increased it in females and decreased it in males. A₁-A_2ARKO mice had lower Temp than WT mice. LA was unaltered in A₁RKO mice and lower in A_2ARKO and A₁-A_2ARKO mice than in WT mice. Caffeine injection increased LA but only in mice expressing A_2AR. Caffeine ingestion also increased LA in an A_2AR-dependent manner in male mice. Caffeine ingestion significantly increased O₂C in WT mice, but less in the different KO mice. Injection of 30 mg/kg caffeine decreased Temp, especially in KO mice, and hence in a manner unrelated to A₁R or A_2AR blockade. Selective A_2B antagonism had little or no effect. Thus A₁R and A_2AR influence HR, Temp, LA, and O₂C in mice in a sex-dependent manner, indicating effects of endogenous adenosine. The A_2AR plays an important role in the modulation of O₂C and LA by acute and chronic caffeine administration. There is also evidence for effects of higher doses of caffeine being independent of both A₁R and A_2AR.

caffeine (1,3,7-trimethylxanthine), the most widely consumed stimulant in the world, acts as a competitive antagonist of adenosine receptors (ARs) (18). It is much more potent on three of the known receptors, A₁, A_2A, and A_2B (A₁R, A_2AR, and A_2BR, respectively) than on the fourth, the A₃R (17). Since a competitive antagonist will have effects only when and where receptors are activated by an agonist, it is relevant that endogenous agonist adenosine is more potent on A₁R and A_2AR than on A_2BR. The potency of adenosine as an agonist in addition depends on the density of the receptors (16). The levels of adenosine are low under physiological condition but are sufficient to partially activate A₁R, A_2AR, and A₃Rs where the receptors are abundantly expressed. For these reasons caffeine is generally believed to act on A₁R and A_2AR to exert most of its effects, at least when given in lower doses (18). However, a recent study showed that A_2BR knockout (A_2BRKO) mice have elevated cytokines, vascular adhesion molecules, and leukocyte adherence (49), indicating that A_2BR is probably also activated by endogenous adenosine. This could indicate that at some sites local adenosine concentrations may be high enough to activate A_2BR.

One way to determine the correctness of this assumption is to examine the effect of caffeine in mice that lack ARs. With the use of this approach, it has been shown that mice in which the A_2AR has been deleted respond less or not at all to caffeine with regard to locomotion and wakefulness (13, 26, 31). Mice that lack A₁Rs, however, show little alteration in locomotion and sleep responses to caffeine (23, 26). In the present study, we have examined caffeine effects in mice that lack either A₁R or A_2AR or both. We also examined effects of the selective A_2BR antagonists, enprofylline (20, 39) and 1-propyl-8-p-sulfophenylxanthine (PSB1115) (11, 22, 48). The parameters we examined were heart rate (HR), body temperature (Temp), locomotor activity (LA), and oxygen consumption (O₂C) since they could be examined in awake animals and for prolonged periods of time. Acute or single-dose caffeine treatment mainly showed a stimulating effect on HR but clearly biphasic effects on Temp and LA (18, 27, 36, 51). We examined not only acute effects of injecting caffeine but also the effect of adding caffeine to the drinking water for 7 days. We have also determined whether the genotype per se leads to changes, since that would indicate an important role of endogenous adenosine in the physiological regulation. Since a previous study (51) revealed major sex differences, we examined both male and female mice.

MATERIAL AND METHODS

Animals.

The study was performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH) and the ethical guidelines of the European Union and Sweden and was approved by the Animal Ethics Committee of Northern Stockholm.

The A₁RKO mice were generated by inactivating the second protein coding exon of the mouse A₁R gene as previously described (29). A_2ARKO mice were generated by inactivating the exon 2 of mouse adenosine A_2AR gene as described previously (8). The A₁RKO mice were cross-bred to C57BL/6 mice for six generations by the Jackson Laboratory using a speed congenic approach, whereas A_2ARKO mice were backcrossed to C57BL/6 mice for more than 10 generations. DNA marker analysis confirmed that >99% of microsatellite DNA markers are from C57BL/6 genetic background in both A₁RKO and A_2ARKO mice. These congenic A₁RKO and A_2ARKO mice were cross bred to generate double heterozygous A₁R-A_2ARKO mice (A₁R+/−, A_2AR+/−). The double heterozygous A₁R-A_2ARKO mice were cross bred to generate all four genotypes from the same littermates, i.e., WT (A₁R+/+, A_2AR+/+), A₁RKO (A₁R−/−, A_2AR+/+), A_2ARKO (A₁R+/+, A_2AR−/−) and A₁R-A_2AR double KO (A₁R−/−. A_2AR−/−) mice. Genotypes were identified by PCR analysis as described previously (8, 29, 44). Mice were kept in individual cages under normal conditions with a 12-h:12-h light-dark cycle during the experiments. All animals were given free access to both food and tap water as described in more detail below (see The telemetry and Oxymax study).

Implantation of telemetry transmitter.

To prevent the infection of mice from the operation, all the instruments and transmitters were sterilized before used in the surgery. The protocol of surgery was described previously (51). Briefly, under anesthesia induced and maintained by isoflurane, the transmitter (TA10ETA-F20) was implanted into the peritoneal cavity and the leads were sutured in a lead II position. Analgesia was ascertained by the local application of 5% xylocaine salve (AstraZeneca, Södertälje, Sweden) and subcutaneous injection of Temgesic (0.3 mg kg⁻¹; Schering-Plough Europe, Brussels, Belgium) after the implantation. Each mouse was allowed at least 7 days of recovery before the start of registration. There was not clear evidence for the infection of mouse after the implantation.

The telemetry and Oxymax study.

We used a telemetry system (Data Sciences, St. Paul, MN) consisting of implantable transmitters (TA10ETA-F20), telemetry receivers (DSI PhysioTel Receivers-RPC-1 Model), and eight universal adapters (UA 10 PC) as previously described (51). A computer program (Data quest A.R.T gold acquisition) sampled calibrated values of HR and Temp as well as noncalibrated LA counts. HR and Temp are recorded as such, whereas LA was determined from recorded displacements of the telemetric device relative to the bottom plate of the cage. Minor movements are not recorded. The protocol is shown in Fig. 1G. After 5 days of baseline registration, the mice received saline, 7.5 mg/kg and 30 mg/kg caffeine (Sigma, St. Louis, MO; dissolved in saline) or enprofylline (kindly provided by Dr. C. G. A. Persson; a selective A_2BR antagonist, in a separate experiment) intraperitoneally injected (bolus) to the mice at 2-h intervals. In yet another separate experiment [using only male wild-type (WT) mice], a single dose of 1-propyl-8-p-sulfophenylxanthine (PSB1115, a highly selective A_2BR antagonist kindly provided by Dr. Christa E. Muller, University of Bonn) was given. After 2 days recovery from the injection, the animals were moved to Oxymax (v 5.11) system one day for the recording of the basal oxygen consumption (O₂C) from changes in total oxygen tension in a closed box. Thereafter, the drinking water was replaced by tap water containing 0.3 g/l caffeine. Again, after 7 days of registration on telemetry, the animals were moved to Oxymax one day for recording of the O₂C. After 8 days of caffeine drinking, the water was changed back to normal tap water. Two days later, the mice were administered a single dose of an unselective β-adrenergic blocker (timolol, 1 mg·kg⁻¹; Sigma).

Fig. 1.

A₁ and A_2A adenosine receptors (A₁R or A_2AR, respectively) affect heart rate (HR) in sex-dependent manner. A and B: HR is recorded for consecutive 10-min periods in wild-type [WT; Female (F), n = 8; Male (M), n = 8], A₁R-knockout (A₁RKO; F, n = 8; M, n = 8), A_2ARKO (F, n = 8; M, n = 10), and A₁-A_2ARKO (F, n = 8; M, n = 9) mice. The dark period is from 7 PM to 7 AM. C and D: summarizing the corresponding top panels. The value for Day shows the HR average between 9 AM and 5 PM; the value for Night shows the HR average between 9 PM and 5 AM; the value for Peak shows the HR average between 7:30 PM and 7:40 PM. In A–D, data are shown as a 5-day average; data are presented as mean values or mean values ± SE (*P < 0.05 compared with WT mice; #P < 0.05 compared with A₁RKO mice; **P < 0.05 compared with A_2ARKO mice; ##P < 0.05 compared with A₁-A_2ARKO mice at the corresponding time period). E: HR was measured telemetrically during 24 h. The data are shown as a 5-day average and presented as mean values or mean values ± SE (*P < 0.05 compared with corresponding female mice). F: effects of saline and timolol (1 mg/kg) on HR in male mice. The HR was measured at 10-min intervals until 80 min after injection. Data are presented as means ± SE (*compared with WT, P < 0.05; #compared with saline injection, P < 0.05). Abbreviations and mice number are the same as above. G shows the protocol of experiments. bpm, Beats/min; NS, saline.

Statistical analysis.

Two-way ANOVA or Student's t-test was employed to evaluate the differences between groups. Data are presented as means ± SE. Statistical significance was defined as P < 0.05.

RESULTS

ARs affect heart rate (HR) in a sex-dependent manner.

HR shows marked diurnal variation (Fig. 1, A and B). At nighttime genotypes (Fig. 1, A and B). Since the lights were progressively dimming, there was a major increase (about 100 beats/min) in HR in both sexes and in all genotypes. During nighttime, HR gradually returned to the daytime values. When the light was switched on, only a small farther fall in HR was observed. Sex difference was marked: HR in male WT mice was about 550 beats/min and was about 60 beats/min higher in female WT mice (Fig. 1E). A similar trend was found in all genotypes, but in A₁RKO mice, the sex difference of HR was less pronounced than in the other mice (Fig. 1E).

There were also clear effects of genotype, but as noted before in the case of A₁R (51), they were partly sex dependent (Fig. 1, A–E). Although HR was similar in female A₁RKO and female WT mice (Fig. 1, A and C), HR was higher in male A₁RKO mice than in male WT mice (Fig. 1, B and D). There was a highly significant decrease in HR (about 35 beats/min) in A_2ARKO mice of both sexes, and HR in A₁-A_2ARKO was also (about 25 beats/min) lower than that in their WT controls of both sexes (Fig. 1, A–E). Again, HR in A₁-A_2ARKO was higher than in A_2ARKO male mice. Thus A₁R and A_2AR appear to have independent and opposite effects on HR in mice.

To determine whether the differences in HR are due to intrinsic differences in the heart or to differences in the autonomic vasomotor control, we administered a single dose of an unselective β-blocker (timolol; 1 mg/kg). This decreased HR in all genotypes compared with a saline injection (Fig. 1F). After the saline injection, the HR was higher in A₁RKO, whereas lower in A_2ARKO and A₁-A_2ARKO than in WT mice. Thus the pattern was similar to that already described in the unstressed animals. Administration of timolol eliminated the HR difference between genotypes in male (Fig. 1F) and female mice (not shown). These data indicated that A₁R and A_2AR regulate HR via altering sympathetic tone.

Acute administration of caffeine (7.5 and 30 mg/kg) had little effect on HR (results not shown) as reported earlier (51). Injection of enprofylline (A_2BR antagonist, at the doses of 7.5 and 30 mg/kg) increased HR in male WT mouse from 529 ± 23 to 590 ± 20 and 562 ± 20 after the low and high dose, respectively (means ± SE; n = 6), but not in females (data not shown). PSB1115 (10 mg/kg) had no effect (PSB1115 vs. saline: 468 ± 16 vs. 479 ± 24, n = 4) in male mice. Week-long oral intake of caffeine decreased HR in WT mice including both sexes, which is more pronounced in the female mouse (50), but there was no clear HR effect in the other genotypes mice (results not shown).

ARs affect body temperature (Temp) in a sex-dependent manner.

Essentially, Temp exhibited similar variations as HR (Fig. 2, A and B). Temp is about 1°C higher in nighttime than in daytime, and this difference was observed in all genotypes. When the lights were turning off, Temp rapidly rose more than 1°C, and then it decreased gradually but did not return to daytime values as long as lights were off; rather, it decreased sharply by 0.5–1°C when the lights were switched on. During daytime, Temp in male WT mice was about 36°C and was about 0.3–0.4°C higher in female mice. During nighttime, the sex difference remained even though the Temp was higher than that during daytime. The sex difference in Temp (Fig. 2E) was less pronounced in A₁-A_2ARKO than in WT mice (and A₁RKO and A_2ARKO mice).

Fig. 2.

A₁R and A_2AR affect body temperature (Temp) in sex-dependent manner. A and B: abbreviations and mice number are the same as in legend to Fig. 1. Temp is recorded for consecutive 10-min periods in mice. The dark period is from 7 PM to 7 AM. C and D: summarizes the corresponding top panels and periods are as described in Fig. 1. The data are shown as a 5-day average and presented as mean values or mean values ± SE (*P < 0.05 compared with WT mice; #P < 0.05 compared with A₁RKO mice; **P < 0.05 compared with A_2ARKO mice; ##P < 0.05 compared with A₁-A_2ARKO mice at the corresponding time period). E: Temp was measured telemetrically during 24 h and recorded for every consecutive 30-min period. The data are shown as a 5-day average and presented as mean values or mean values ± SE (*P < 0.05 compared with corresponding female mice). F: injection of saline (NS), caffeine/enprofylline (Enp) 7.5 mg/kg or 30 mg/kg, was administered intraperitoneally at 2-h intervals in WT1 (F, n = 5; M, n = 6), WT2 (F, n = 8; M, n = 8), A₁RKO (F, n = 8; M, n = 8), A₁-A_2ARKO (F, n = 8; M, n = 9), and A_2ARKO mice (F, n = 8; M, n = 10). Temp is shown as the average from 70 to 80 min after the injection. Data are presented as means ± SE (*P < 0.05 compared with saline injection).

Deletion of the A₁Rs had small effects on Temp (Fig. 2, A and B). By contrast, deletion of the A_2ARs had a major effect, especially in the daytime and surprisingly the effect was qualitatively different in males and females (Fig. 2, A–D). Thus in females it was higher (Fig. 2A), whereas in males it was lower (Fig. 2, B and D). The Temp of the double KO mice tended to decrease in both sexes (Fig. 2, A–D). Female mice had higher Temps (about 0.2 to 0.7°C) than males (Fig. 2E) in all genotypes, and the difference was most pronounced in A_2ARKO mice and least in A₁-A_2ARKO mice.

Acute administration of caffeine had a biphasic effect on Temp (Fig. 2F): a dose of 7.5 mg/kg caffeine increased Temp in all genotypes, whereas the administration of 30 mg/kg caffeine produced a smaller effect in WT mice and actually decreased Temp compared with saline injection in the other genotypes. A high dose of enprofylline (30 mg/kg) also decreased Temp compared with saline injection in female (but not in males) WT mice as studied in a separate experiment (Fig. 2F), but a low dose (7.5 mg/kg) enprofylline had little effect on mouse Temp. The selective antagonist PSB1115 did not change Temp relative to a saline injection (at 70 min after injection: saline gave a Temp of 36.09 ± 0.34°C and after PSB1115 injection the Temp was 36.23 ± 0.14°C; n = 4). As with HR, chronic oral intake of caffeine had little effect on Temp (data not shown).

ARs affect locomotor activity and O₂C in a sex-dependent manner.

A pronounced diurnal pattern is also seen in LA of the mice (Fig. 3, A–D). LA in nighttime is about 2–5 times higher than in daytime, and this difference was observed in all genotypes. When the lights were turning off, the mice started moving around quickly (up to 16 counts/min in female WT mice); then their activity gradually declined (from 16 to 6 counts/min in female WT mice and from 10 to 5 in male mice). When the lights were turning on, the mice adopted the daytime activity pattern. The sex difference of LA was much more pronounced in the dark than that in the light (Fig. 3, A–D). The female mice were more active (2 to 3 counts/min higher) than the males (Fig. 3E).

Fig. 3.

A₁R and A_2AR affect locomotor activity (LA) in sex-dependent manner. Graphs constructed essentially as described in legend to Fig 2. A and B: LA is recorded for consecutive 30-min periods in mice. The dark period is from 7 PM to 7 AM. C and D: summarizes the corresponding top panels. Mean values or mean values ± SE (*P < 0.05 compared with WT mice; #P < 0.05 compared with A₁RKO mice; **P < 0.05 compared with A_2ARKO mice; ##P < 0.05 compared with A₁-A_2ARKO mice at the corresponding time period) are given. E: LA were measured telemetrically during 24 h and recorded for every consecutive 30-min periods. The data are shown as a 5-day average and presented as mean values or mean values ± SE (*P < 0.05 compared with corresponding female mice). F: normal saline (NS), caffeine/enprofylline (Enp) 7.5 mg/kg or 30 mg/kg, was administered intraperitoneally at 2-h intervals in WT1 (F, n = 5; M, n = 6), WT2 (F, n = 8; M, n = 8), A₁RKO (F, n = 8; M, n = 8), A₁-A_2ARKO (F, n = 8; M, n = 9), and A_2ARKO mice (F, n = 8; M, n = 10). LA is shown as the average from 60 to 90 min after the injection. Data are presented as means ± SE (*P < 0.05 compared with saline injection).

Locomotor activity was essentially unaltered in A₁RKO but lower in A_2ARKO and A₁-A_2ARKO than in WT mice (Fig. 3, A and B). In A₁-A_2ARKO mice, the increase in activity normally seen when the lights are dimming was markedly reduced (Fig. 3, A–D).

As expected, caffeine caused a dose-dependent stimulation of locomotion in WT mice, which was markedly reduced in the other genotypes and even eliminated in mice lacking A_2ARs (Fig. 3F). However, injection of the A_2B antagonist enprofylline did not affect the LA in WT in another set of experiment (Fig. 3F). PSB1115 (10 mg/kg) was similarly ineffective (not shown).

Chronic caffeine ingestion did increase LA in A₁RKO, mice and this was mainly observed in nighttime (Fig. 4, C and D). LA was also increased in WT and A₁-A_2ARKO male mice, but no significant change was observed in the corresponding females (Fig. 4, A, B, E, and F). There was no difference in LA in A_2ARKO mice after caffeine ingestion compared with before caffeine ingestion.

Fig. 4.

Effects of chronic caffeine intake on LA and oxygen consumption (O₂C) in mice. After the baseline recording (Water), mice drank tap water with 0.3 g/l caffeine for 7 days (Caffeine). The LAs in the water group are shown as 5-day averages in 30-min intervals, whereas LA data in caffeine group are shown as 3-day averages (days 5, 6, and 7 during chronic caffeine treatment). O₂C is shown as 1-h average. The values are presented as means ± SE (*Compared with Water, P < 0.05). Abbreviations are the same as in legend Fig. 2 or otherwise indicated. A and B: chronic caffeine ingestion increased oxygen consumption in WT mice (Water: F, n = 8; M, n = 8; Caffeine: F, n = 7, M, n = 8). It also increased LA in male WT mice but not in female WT mice (Water: F, n = 7; M, n = 6; Caffeine: F, n = 9, M, n = 6). C and D: in A₁RKO mice, chronic caffeine ingestion increased LA in both females and males (Water: F, n = 8; M, n = 8; Caffeine: F, n = 8, M, n = 8) but had no effect on O₂C (Water: F, n = 7; M, n = 9; Caffeine: F, n = 9, M, n = 8). E and F: in A₁-A_2ARKO mice, chronic caffeine ingestion increased LA in males but not in females (Water: F, n = 8; M, n = 10; Caffeine: F, n = 8, M, n = 9). Chronic caffeine ingestion did not affect O₂C (Water: F, n = 6; M, n = 7; Caffeine: F, n = 6, M, n = 7). G and H: in A_2ARKO mice, chronic caffeine ingestion increased O₂C in males but not in females (Water: F, n = 7; M, n = 6; Caffeine: F, n = 7, M, n = 4). LA was not affected in females or males (Water: F, n = 8; M, n = 10; Caffeine: F, n = 8, M, n = 8).

As for the other parameters studied, there was a velar diurnal pattern in oxygen consumption (O₂C). This could be observed despite the fact that for technical reasons this was measured mainly during the nighttime (Fig. 5, A and B). Sex difference in O₂C was also revealed (Fig. 5C), and the female mice consumed more oxygen than the male mice.

Fig. 5.

Genotype differences in O₂C. Abbreviations are the same as in legend to Fig. 1. A: O₂C was measured for 60-min periods in WT (F, n = 7; M, n = 6), A₁RKO (F, n = 7; M, n = 9), A₁-A_2ARKO (F, n = 6; M, n = 7), and A_2ARKO mice (F, n = 7; M, n = 4) mice. The dark period is from 7 PM to 7 AM. B summarizes the top panel. The black column shows the O₂C average between 9 PM and 5 AM; the spotted column for Peak shows the O₂C average between 7:30 PM and 8:30 PM. Data are presented as mean values or mean values ± SE (*P < 0.05 compared with WT mice). C: O₂C (Oxygen) was measured for 16-h mice.

There was no major difference in O₂C between genotypes in the females (data not shown). In male mice (Fig. 5, A and B), the O₂C in A_2ARKO and A₁-A_2ARKO male mice was lower than in the WT controls.

Chronic caffeine ingestion significantly increased O₂C in WT mice (Fig. 4, A and B). In A_2ARKO male, mice oxygen consumption tended to increase, but this was not the case in females (Fig. 4, G and H) or in the A₁RKO and A₁-A_2ARKO mice. When comparing with the caffeine effects on LA, the increase of O₂C was not always consistent with the increase of LA via chronic caffeine ingestion.

DISCUSSION

The present study provides some evidence that endogenous adenosine acting on A₁R and/or A_2ARs influences HR, Temp, locomotion and O₂C under physiological conditions. We will briefly discuss these possible roles.

Regulation of HR by ARs.

All four ARs are present in the mouse heart (51). However, the mRNA levels of A₁ and A_2A are about 100-fold higher than those of A_2BR and A₃R (51), and the affinity of adenosine to A₁R and A_2AR is at least as high as to A_2B and A₃Rs (17). Since the potency of adenosine as an agonist depends on the density and affinity of the receptors (16, 17), the possibility exists that intracardiac A₁R and A_2AR could regulate HR. However, indirect effects are also possible. Our data indicate that both A₁R and A_2AR were involved in physiological HR regulation. A₁RKO mice had higher HR than WT mice in males and A₁-A_2ARKO mice had higher HR than A_2ARKO mice in both females and males, indicating that endogenous adenosine lowered HR via the A₁R under physiological conditions. This confirms our previous study showing that isolated hearts from A₁RKO mice had higher HR than those from WT controls (51). It is well known that the activation of the A₁R decreases HR by slowing atrioventricular nodal conduction (4, 25), and adenosine is used clinically to alleviate supraventricular tachyarrhythmia. Surprisingly, although the HR of male A₁RKO mice was higher than that of WT mice, the HR of female A₁RKO mice was not increased. This indicates an interesting sex difference in A₁R function of regulating HR. It may be related to the higher HR in females, but the exact reason for the sex difference remains to be explored in future studies.

In this study, we also found that A_2ARKO mice had significantly lower HR than WT mice, and HR of A₁-A_2ARKO mice was much lower than that of A₁RKO mice, suggesting that adenosine increases HR via activation of A_2AR. This was surprising since a selective A_2AR agonist does not influence HR of isolated hearts (33), indicating that the A_2AR plays a minor role in local regulation of HR. However, systemic administration of selective A_2AR agonists can increase HR, probably due to reflex activation secondary to reduced blood pressure (6, 31, 32, 34, 35). In addition, there is evidence that there are ARs in the brain stem regulating cardiovascular control centers (2–4).

Given that A_2ARs apparently mediate a positive effect and A₁Rs a negative effect on HR, it was interesting that the HR in A₁-A_2ARKO mice was lower than in WT controls. This suggests that under physiological conditions, the influence of the A_2AR on HR is actually larger than that of the A₁R, despite the fact that the latter effect is much more widely recognized.

We found that the administration of timolol (β-blocker) completely abolished the difference in HR between ARKO and WT mice. This argues that the endogenous adenosine that regulates HR under physiological conditions is acting not mainly on A₁R and A_2AR on cardiac cells but by regulating nervous input [see reviews (5, 15, 17)]. This could be achieved in many ways. Adenosine, through the activation of A₁R, decreases the release of neurotransmitters including norepinephrine and acetylcholine from sympathetic and parasympathetic neurons (19) and via A₁R and A_2AR regulates firing of important cell groups in central nervous system and directly influences cardiovascular control centers (1, 2). Furthermore, adenosine acting at chemoreceptors (14) or on cells in the carotid body (30) could cause reflex activation of sympathetic nerves. Finally, the HR changes may be secondary to changes in blood pressure. However, the A_2ARKO mice show no change in blood pressure (40) and A₁RKO mice show either a small increase (7) or no change (42). Thus alterations in blood pressure cannot readily explain the observed HR differences between genotypes. Similarly a major increase in pressure after β-receptor blockade by timolol is unlikely as an explanation for the fall in HR.

Regulation of Temp by adenosine receptors.

Adenosine (47) and other AR agonists (10, 43) can induce profound hypothermia, indicating that AR are involved in the regulation of Temp. It has been shown that a selective A₁R agonist (CPA) decreases Temp (43), and this effect is reduced in mice lacking A₁Rs (25) and improves tolerance to cold temperature (47). We found the Temp of female A₁RKO mice to be slightly higher than in WT mice, but this was not observed in male A₁RKO mice. Thus, endogenous adenosine acting on A₁Rs appears to have a modest influence on Temp in unstressed mice.

The role of A_2AR in the regulation of Temp is controversial. It was reported that the administration of a selective A₂R agonist did not influence Temp (43), but at the time agonists were not very selective. However, in a later study a selective A_2AR agonist decreased Temp in the rat (9). In the present study, we found that A_2ARs clearly appeared to play a role in regulating Temp, but this was strongly sex dependent. Although female A_2ARKO mice had higher Temp than WT mice, the Temp of male A_2ARKO mice was lower than that of WT mice. This difference was particularly obvious in daytime. We lack an explanation for this observation but can conclude that neither differences in activity nor differences in metabolic rate as judged by O₂C appear to provide a complete explanation. We also do not ascribe the difference to the estrous cycle, since we averaged the temperature over 5 days and since variability was similar in males and females. The activation of A₁R and especially A_2AR by adenosine either in central never system (38) or locally (41) could induce a cutaneous vasodilatation, subsequently increased heat loss, and reduced Temp. It is, however, difficult to see how this could adequately explain the qualitative differences between males and females.

Regulation of locomotor activity and O₂C by A₁R or A_2AR.

The roles of the different ARs in regulating activity state including LA are still incompletely understood. We found LA to be slightly higher in female (but not male) A₁RKO mice than in their WT controls, but the effect is minor. This small effect agrees with several previous studies on A₁KO mice (18, 25). A_2ARKO mice had lower LA than their WT controls, implying A_2AR stimulated LA. Although LA response to the light turning off in A₁-A_2ARKO mice was smaller than in A_2ARKO mice, there was no difference in LA between A_2ARKO and A₁-A_2ARKO mice. These findings suggest that A_2ARs play much more important roles than A₁Rs in regulating LA.

LA in A₁-A_2ARKO mice was much lower than in WT mice, which is in apparent contrast with the finding that pharmacological blockage of A₁R and A_2AR elevated locomotor activities in rats (28). Similarly, several studies have indicated a small stimulatory effect of A_2AR antagonists, but A_2ARKO mice have lowered activity. The difference could be partially related to the difference between acute blockade and the life-long elimination of a receptor and perhaps also to the difference between eliminating the receptor and partially blocking it with a competitive antagonist. The latter possibility is supported by our recent finding that double heterozygous mice show increased LA (50).

Male A_2ARKO and A₁-A_2ARKO mice consumed less oxygen than male WT mice. This is consistent with their lower LA. However, we did not find clear differences in O₂C among the females, suggesting that oxygen demand is not always covariant with spontaneous activity and that the regulation of oxygen demand via ARs is also sex dependent.

Caffeine effects.

The roles of the different ARs can be revealed also by the use of caffeine as an antagonist. As expected, the administration of caffeine, either as a single dose or repeatedly in the drinking water, increased the activity of mice as determined by measuring locomotor activity. The effect of the single dose was dose dependent. The effect of the oral caffeine intake was not observed in the female mice but only in the male WT mice, perhaps because they showed a lower basal activity than the females. The stimulating effect of single injections of caffeine was not clearly altered in the A₁KO mice of either sex, but the nighttime activity appeared to be more strongly affected by oral caffeine in these mice than in WT mice. As also expected from previous data, effects of caffeine were largely eliminated in mice lacking A_2AR. Caffeine, injected or ingested, had a statistically significant stimulatory effect in male mice, but it was small. These results suggested that single-dose caffeine injection and chronic caffeine intake increased LA via actions that depended more on A_2AR than on A₁R.

Caffeine blocks not only A₁ and A_2AR but also A_2BR. We therefore compared the effect of caffeine with that of enprofylline, which is about equipotent with theophylline and hence more potent than caffeine as an A_2B antagonist (18, 20, 39). It was chosen in preference to some more recent and selective antagonists because its pharmacokinetics is well known and direct in vivo comparisons with the parent compound have been repeatedly performed (37, 45). It essentially lacks activity at other ARs. Enprofylline and PSB1115 did not activate the mice, and we conclude that A₁ and A_2AR are the important targets for this action. Moreover, as discussed in detail elsewhere (18, 21), the basal gangla is an important target area.

HR was scarcely affected by caffeine essentially as described (51). Enprofylline had a minor effect in male but no effect in female mice. Enprofylline is not only an adenosine A_2BR antagonist but also acts as a phosphodiesterase inhibitor (45) and could by this means increase HR (24). It is also virtually equipotent with caffeine on most studied isoforms of PDE (45) and less potent than other caffeine metabolites such as theophylline or paraxanthine. The more specific A_2B antagonist PSB1115 (11, 12, 22, 48) had no effect on HR, suggesting that induced PDE inhibition may be important. We also did not observe any changes in Temp following long-term oral caffeine ingestion. By contrast, acute injection of the lowest dose (7.5 mg/kg) of caffeine caused a small temperature increase. A higher dose of caffeine (30 mg/kg) had little effect on Temp in WT mice but caused a clear-cut decrease in the other genotypes. The fall was quite dramatic. This clearly indicates a biphasic effect of caffeine on temperature and that this depressant effect does not require either A₁ or A_2AR. Since enprofylline caused at most a modest fall in body temperature, and PSB1115 (10 mg/kg) had no effect at all, we conclude that the temperature lowering effect of 30 mg/kg caffeine in the KO mice is most likely unrelated to AR blockade. The modest fall is compatible with some role of PDE inhibition, especially as we find that a very selective PDE inhibitor (rolipram) had a temperature lowering effect that was, as in the case of enprofylline, especially obvious in male mice (Yang and Fredholm, unpublished observations).

It is well recognized that caffeine can have biphasic effects on, e.g., locomotion and mood exhibiting stimulation at low doses and inhibitory effects at higher doses (18). In a study of A_2ARKO mice, it was shown that the stimulatory effect on activity required A_2ARs (13) as confirmed here, but the authors suggested that the depressant effect, which remained, was perhaps due to A₁R blockade. In a follow-up study in our laboratory, the first conclusion was supported, but the second was not because in A₁RKO mice the depressant effect also remained (23). The present finding of a major fall in Temp at doses where this depressant effect is commonly recorded provides an intriguing possible explanation and also indicates that we are dealing with an effect that is not dependent on either A₁ or A_2AR.

In conclusion, both A₁ and A_2AR, which are activated by endogenous adenosine under normal physiological conditions, are involved in the regulation of HR, Temp, locomotion activities, and oxygen demand in mice in a sex-dependent manner. The regulation of HR by adenosine through A₁R and A_2AR is mainly due to altering the sympathetic tone. The A_2AR plays a very important role in the modulation of oxygen demand and locomotor activities by acute and chronic caffeine administration.

GRANTS

This study was supported by the Swedish Science Council Grant No. 2553, by NIH Grant RO1 NS048995, the Swedish Heart and Lung Foundation, Karolinska Institutet, and the European Commission (CT 2005-518189). The content of this paper is the responsibility of the authors, and the granting agencies take no responsibility therefore.